1. Field of the Invention
The present disclosure relates to a depth generating method, more particularly, to an image capture device, a depth generating device and a method thereof capable of combining two depth calculation technologies to improve precisions of the calculated depths.
2. Description of the Related Art
Common depth generating technologies can be classified into two types. The first technology is performed by a camera device cooperated with a pattern projection device. The pattern projection device projects a structure light with a preset pattern to an external environment and the camera device then captures an image from the external environment. Because the objects in the external environment are not located at the same plane, the image captured by the camera device will contain a deformed preset pattern. The depth of each object in the captured image can be calculated according to the deformation of the preset pattern in the captured image.
However, the first technology has a drawback in low resolution of generated depths due to the limitation of the structure light. At present, the amount of pixels of the camera on the hand-held device becomes higher and higher, so the drawback of low resolution limits application scope of the first technology. In addition, the first technology cannot be applied in the external environment with strong ambient light, because the structure light is not clear enough for calculating the depths.
The second technology is to capture two images with different view angles from the external environment, for example, one camera device is used to capture images at different positions, or two camera devices capture images at the same time. Next, the two images are performed image process to search corresponding feature points, and depths of each of objects in the images are calculated according to these feature points. The second technology can calculate depth of each of pixels in the image, so the resolution of calculated depths becomes higher correspondingly while the amount of pixels of the camera becomes higher. Besides, the second technology is not affected by the ambient light of the external environment.
However, the second technology has a drawback in that if the image contains single color block or repeated pattern, it is easy to cause the misjudgment in comparison of feature points between the two images, so the precisions of calculated depths are decreased correspondingly.
Therefore, what is need is a depth generating method which is not affected by the ambient light and can calculate depths with high resolution and high precisions.